Arterial blood gases fail to reflect acid-base status during cardiopulmonary resuscitation: a preliminary report

Point-of-care laboratory analyses of intraosseous, arterial and central venous samples during experimental cardiopulmonary resuscitation

INTRODUCTION

Screening and correcting reversible causes of cardiac arrest (CA) are an essential part of cardiopulmonary resuscitation (CPR). Point-of-care (POC) laboratory analyses are used for screening pre-arrest pathologies, such as electrolyte disorders and acid-base balance disturbances. The aims of this study were to compare the intraosseous (IO), arterial and central venous POC values during CA and CPR and to see how the CPR values reflect the pre-arrest state.

METHODS

We performed an experimental study on 23 anaesthetised pigs. After induction of ventricular fibrillation (VF), we obtained POC samples from the IO space, artery and central vein simultaneously at three consecutive time points. We observed the development of the values during CA and CPR and compared the CPR values to the pre-arrest values.

RESULTS

The IO, arterial and venous values changed differently from one another during the course of CA and CPR. Base excess and pH decreased in the venous and IO samples during untreated VF, but in the arterial samples, this only occurred after the onset of CPR. The IO, arterial and venous potassium values were higher during CPR compared to the pre-arrest arterial values (mean elevations 4.4 mmol/l (SD 0.72), 3.3 mmol/l (0.78) and 2.8 mmol/l (0.94), respectively).

CONCLUSIONS

A dynamic change occurs in the common laboratory values during CA and CPR. POC analyses of lactate, pH, sodium and calcium within IO samples are not different from analyses of arterial or venous blood. Potassium values in IO, arterial and venous samples during CPR are higher than the pre-arrest arterial values.

Severe metabolic acidosis after out-of-hospital cardiac arrest: risk factors and association with outcome

BACKGROUND

Metabolic acidosis is frequently observed as a consequence of global ischemia-reperfusion after out-of-hospital cardiac arrest (OHCA). We aimed to identify risk factors and assess the impact of metabolic acidosis on outcome after OHCA.

METHODS

We included all consecutive OHCA patients admitted between 2007 and 2012. Using admission data, metabolic acidosis was defined by a positive base deficit and was categorized by quartiles. Main outcome was survival at ICU discharge. Factors associated with acidosis severity and with main outcome were evaluated by linear and logistic regressions, respectively.

RESULTS

A total of 826 patients (68.3% male, median age 61 years) were included in the analysis. Median base deficit was 8.8 [5.3, 13.2] mEq/l. Male gender (p = 0.002), resuscitation duration (p < 0.001), initial shockable rhythm (p < 0.001) and post-resuscitation shock (p < 0.001) were associated with an increased level of acidosis. ICU mortality rate increased across base deficit quartiles (39.1, 59.2, 76.3 and 88.3%, p for trend < 0.001), and base deficit was independently associated with ICU mortality (p < 0.001). The proportion of CPC 1 patients among ICU survivors was similar across base deficit quartiles (72.8, 67.1, 70.5 and 62.5%, p = 0.21), and 7.3% of patients with a base deficit higher than 13.2 mEq/l survived to ICU discharge with complete neurological recovery.

CONCLUSION

Severe metabolic acidosis is frequent in OHCA patients and is associated with poorer outcome, in particular due to refractory shock. However, we observed that about 7% of patients with a very severe metabolic acidosis survived to ICU discharge with complete neurological recovery.

Improving standard cardiopulmonary resuscitation with an inspiratory impedance threshold valve in a porcine model of cardiac arrest

To improve the efficiency of standard cardiopulmonary resuscitation (CPR), we evaluated the potential value of impeding respiratory gas exchange selectively during the decompression phase of standard CPR in a porcine model of ventricular fibrillation. After 6 min of untreated cardiac arrest, anesthetized farm pigs weighing 30 kg were randomized to be treated with either standard CPR with a sham valve (n = 11) or standard CPR plus a functional inspiratory impedance threshold valve (ITV(TM)) (n = 11). Coronary perfusion pressure (CPP) (diastolic aortic minus right atrial pressure) was the primary endpoint. Vital organ blood flow was assessed with radiolabeled microspheres after 6 min of CPR, and defibrillation was attempted 11 min after starting CPR. After 2 min of CPR, mean +/- SEM CPP was 14 +/- 2 mm Hg with the sham valve versus 20 +/- 2 mm Hg in the ITV group (P < 0.006). Significantly higher CPPs were maintained throughout the study when the ITV was used. After 6 min of CPR, mean +/- SEM left ventricular and global cerebral blood flows were 0.10 +/- 0.03 and 0.19 +/- 0.03 mL. min(-1). g(-1) in the Control group versus 0.19 +/- 0.03 and 0.26 +/- 0.03 mL. min(-1). g(-1) in the ITV group, respectively (P < 0.05). Fifteen minutes after successful defibrillation, 2 of 11 animals were alive in the Control group versus 6 of 11 in the ITV group (not significant). In conclusion, use of an inspiratory impedance valve during standard CPR resulted in a marked increase in CPP and vital organ blood flow after 6 min of cardiac arrest.

A computer model of gas exchange during cardiopulmonary resuscitation

This paper presents a computer model of gas exchange during cardiopulmonary resuscitation (CPR) that permits independent adjustment of inspired air content (16% O2 and 4.5% CO2 present in mouth-to-mouth (MTM) ventilation or ambient air), shunt, deadspace, diffusion impairment, cardiac output, and ventilation. The model contains 15500 acini, each with its own blood supply. Gas exchange occurs at each perfused and ventilated acinus. Arterial P(O2) and P(CO2) are calculated from the summed arterial blood flow using standard formulae. The model and simulations show that MTM ventilation provides inadequate oxygenation when the victim is at high altitude or has diffusion impairment. They also show that analysis of inspired and expired gas concentrations to measure gas exchange primarily measures wash in and wash out of gas when cardiac output is low and that this explains the negative oxygen consumption and carbon dioxide production measured in a previous study.

Sodium bicarbonate in cardiac arrest: a reappraisal

The routine use of sodium bicarbonate in patients with cardiac arrest has been discouraged, with the benefit of outcome evaluation. Current recommendations include an elaborate stratification of circumstances in which bicarbonate is to be used. The physiological and clinical aspects of bicarbonate administration during cardiopulmonary resuscitation in animal and human studies were reviewed. The onset of significant acidemia or alkalemia is associated with adverse system specific effects. The administration of bicarbonate may mitigate the adverse physiological effects of acidemia, improve response to exogenously administered vasopressor agents, or simply increase venous return due to an osmolar effect, resulting in increased coronary perfusion pressure. Likewise, bicarbonate may have adverse effects in each of these areas. The preponderance of evidence suggests that bicarbonate is not detrimental and may be helpful to outcome from cardiac arrest. An objective reappraisal of the empirical use of bicarbonate or other buffer agents in the appropriate "therapeutic window" for cardiac patients may be warranted.

Effect of hemorrhagic shock and reperfusion on the respiratory quotient in swine

OBJECTIVES

Respiratory quotient, the ratio of CO2 production to oxygen consumption (VO2), is principally affected by the fuel source used for aerobic metabolism. Since the respiratory quotient, VO2, and CO2 production cannot be directly measured easily, indirect calorimetry is commonly used to determine the value of these variables at the airway level (i.e., airway respiratory quotient, airway VO2, and airway CO2 production). However, under nonsteady-state conditions, a variety of phenomena can alter the relationship between true metabolic activity and measurements determined by indirect calorimetry. During exercise, for example, airway respiratory quotient increases as anaerobic threshold is reached because of the disproportionate increase in airway CO2 production that results from the CO2 liberated through the buffering of excess hydrogen ions by bicarbonate. We hypothesized that hemorrhage and reinfusion might change airway respiratory quotient in a consistent manner as shock is produced and reversed.

DESIGN

Prospective laboratory study.

SETTING

University animal laboratory.

SUBJECTS

Eight pigs (25 +/- 2 [SD] kg), anesthetized with fentanyl and relaxed with pancuronium bromide, and mechanically ventilated on room air.

INTERVENTIONS

The animals were sequentially hemorrhaged and then autotransfused while metabolic and hemodynamic measurements were obtained, using continuous indirect calorimetry and continuous applications of the Fick principle. Hemoglobin, arterial lactate concentration, and blood gases for calibration were measured serially. Analysis of variance was used to compare various periods in time.

MEASUREMENTS AND MAIN RESULTS

Between baseline and peak hemorrhage, and between peak hemorrhage and postreinfusion, all of the following variables changed significantly (p < .05): airway VO2 (baseline 6.4 +/- 0.9 mL/min/kg, peak hemorrhage 3.9 +/- 0.6 mL/min/kg, postreinfusion 7.0 +/- 1.4 mL/min/kg); airway CO2 production (baseline 5.5 +/- 0.9 mL/min/kg, peak hemorrhage 4.5 +/- 0.9 mL/min/kg, postreinfusion 6.0 +/- 1.4 mL/min/kg); airway respiratory quotient (baseline 0.87 +/- 0.07, peak hemorrhage 1.16 +/- 0.07, postreinfusion 0.87 +/- 0.05); lactate concentration (baseline 2.4 +/- 1.2 mmol/L, peak hemorrhage 6.7 +/- 1.9 mmol/L, postreinfusion 5.1 +/- 2.0 mmol/L); and delta PCO2 (venous PCO2-PaCO2) (baseline 4.5 +/- 3.6 torr [0.6 +/- 0.5 kPa], peak hemorrhage 12.1 +/- 5.3 torr [1.6 +/- 0.7 kPa], postreinfusion 2.7 +/- 2.7 torr [0.4 +/- 0.4 kPa]).

CONCLUSIONS

Airway respiratory quotient increases in hemorrhagic shock and decreases again as shock is reversed during reinfusion. This phenomenon appears related to the buffering of excess of hydrogen ion during hemorrhagic shock.

Changes in arterial and mixed venous blood gases during untreated ventricular fibrillation and cardiopulmonary resuscitation

This investigation was designed to evaluate the changes in arterial and mixed venous acid-base conditions during untreated ventricular fibrillation and after institution of cardiopulmonary resuscitation (CPR). Fifty-two swine (weight: 25-40 kg) were studied after induction of ventricular fibrillation. In a subgroup of 10 animals, 10-min CPR trials were performed. Arterial and mixed venous blood gases were monitored at baseline, after 5 min of untreated ventricular fibrillation (nonintervention interval) and after 10 min of mechanical CPR. Standard CPR was performed at compression rates of 100/min with a 60% duty cycle. Arterial pH, Pco2, and HCO3 were unchanged when baseline values were compared with those obtained after 5 min of untreated ventricular fibrillation, while arterial Po2 decreased from 81 to 69 torr. Mixed venous pH decreased from 7.41 to 7.35, Pco2 increased from 43 to 48 torr, Po2 decreased from 40 to 38 torr and HCO3 decreased from 28 to 26 mEq/l (P < 0.05). Although these changes were statistically significant, many remain in the normal range. Both arterial and mixed venous pH and HCO3 fell further after 9 min of CPR and Pco2 increased (P < 0.05). Alterations in mixed venous pH and Pco2 were more apparent than corresponding changes in arterial blood gas composition. We conclude that untreated cardiac arrest may be accompanied by normal arterial and mixed venous blood gas levels. Tissue acidosis is only revealed after tissue perfusion is restored and is most accurately reflected in the mixed venous blood gas composition. This apparent paradox provides insight into the relationship between tissue perfusion and arterial and mixed venous acid-based composition.

The relationship between the arteriovenous carbon dioxide gradient and cardiac index

It has been reported that under normal conditions, mixed venous blood gases have approximated arterial samples; however, during cardiac arrest or severe cardiogenic shock, marked differences between arterial and venous blood gases have been noted. To further assess the relationships between arterial and mixed venous blood gases and cardiac index, a study population was chosen consisting of patients with less severe states of cardiac impairment. The differences between arterial and mixed venous PCO2s and pHs were compared with cardiac indexes (CI) of 44 patients in an intensive care unit with arterial lines and Swan-Ganz catheters in place. Twenty-six patients with normal CIs (2.6 to 4.1 L/min/m2) had a mean difference in mixed venous-arterial PCO2 (delta PCO2) of 4.88 +/- 0.40 mm Hg. In patients with low CIs (< 2.6), the delta PCO2 was 7.44 +/- 0.63 mm Hg (P = .001). The difference of mixed venous and arterial pH (delta pH) was 0.027 +/- 0.004 pH units for patients with normal CIs and 0.04 +/- 0.003 pH units for those with low CIs (P < .002). When the CIs of all patients were plotted against the delta PCO2s, there was an inverse linear relationship wherein delta PCO2 increased as CI decreased (r = -.47, P = .0011). There is an inverse relationship between delta PCO2 and CI that has not been previously described. An elevated delta PCO2 may be a marker of a low cardiac index.

Bystander cardiopulmonary resuscitation. Is ventilation necessary?

BACKGROUND

Prompt initiation of bystander cardiopulmonary resuscitation (CPR) improves survival. Basic life support with mouth-to-mouth ventilation and chest compressions is intimidating, difficult to remember, and difficult to perform. Chest compressions alone can be easily taught, easily remembered, easily performed, adequately taught by dispatcher-delivered telephone instruction, and more readily accepted by the public. The principal objective of this study was to evaluate the need for ventilation during CPR in a clinically relevant swine model of prehospital witnessed cardiac arrest.

METHODS AND RESULTS

Thirty seconds after ventricular fibrillation, swine were randomly assigned to 12 minutes of chest compressions plus mechanical ventilation (group A), chest compressions only (group B), or no CPR (group C). Standard advanced cardiac life support was then provided. Animals successfully resuscitated were supported for 2 hours in an intensive care setting, and then observed for 24 hours. All 16 swine in groups A and B were successfully resuscitated and neurologically normal at 24 hours, whereas only 2 of 8 group C animals survived for 24 hours (P < .001, Fisher's exact test). One of the 2 group C survivors was comatose and unresponsive.

CONCLUSIONS

In this swine model of witnessed prehospital cardiac arrest, the survival and neurological outcome data establish that prompt initiation of chest compressions alone appears to be as effective as chest compressions plus ventilation and that both techniques of bystander CPR markedly improve outcome compared with no bystander CPR.

Arterial blood gases during cardiac arrest: markers of blood flow in a canine model

Measures of CO2 have been shown to correlate with coronary perfusion pressure and cardiac output during cardiac arrest. We evaluated arterial pH (pHa) relative to blood flow during cardiac arrest in a canine electromechanical dissociation (EMD) model of cardiac arrest using different resuscitation techniques. Following 15 min of cardiac arrest, 24 mongrel dogs received epinephrine with continued CPR or closed-chest cardiopulmonary bypass. Central arterial blood gases, end-tidal carbon dioxide (PetCO2), coronary perfusion pressure and cardiac output were measured. During CPR, prior to epinephrine or bypass, there was no correlation of pHa, PACO2 and PetCO2, with cardiac output or coronary perfusion pressure. Immediately after instituting the resuscitation techniques, both pHa and PaCO2 showed a significant correlation with cardiac output (pHa; R = -0.78, P less than 0.001 and PaCO2; R = 0.87, P less than 0.001) and with coronary perfusion pressure (pHa; R = -0.75, P less than 0.001 and PaCO2; R = 0.75, P less than 0.001). Eventual survivors (n = 15) had an early significant decrease in pHa, base excess and a significant increase in PaCO2 which was not present in non-survivors (n = 9). Neither pHa nor PaCO2 correlate with blood flow under low flow conditions of CPR. However, with effective circulatory assistance, pHa and PaCO2 reflect systemic blood flow and reperfusion washout.

Cardiac effects of carbon dioxide-consuming and carbon dioxide-generating buffers during cardiopulmonary resuscitation

Recent studies have demonstrated an increase in carbon dioxide (CO2) tension (PCO2) in both mixed venous and coronary vein blood early in the course of cardiac arrest and cardiopulmonary resuscitation. Because increased PCO2 in the myocardium correlates with both ischemic injury and depression of contractile function, the effects of hypertonic solutions of either the CO2-"generating" sodium bicarbonate (NaHCO3) buffer, a mixture of sodium carbonate (Na2CO3) and sodium bicarbonate (carbicarb) acting as a CO2-"consuming" buffer, or saline placebo (NaCl) were compared during cardiopulmonary resuscitation in 25 healthy minipigs. Both buffer agents significantly increased the pH and HCO3- of arterial, mixed venous and coronary vein blood. Bicarbonate increased whereas carbicarb reduced blood PCO2 in the systemic circuit as anticipated. However, neither the PCO2 nor the lactate content of coronary vein blood was favorably altered by buffer therapy. Four of eight animals treated with bicarbonate, five of eight treated with carbicarb and six of nine placebo-treated animals were successfully resuscitated and had a comparable 24 h survival rate. Coronary perfusion pressure during precordial compression, a critical determinant of resuscitability, was transiently decreased by each of the hypertonic solutions. Accordingly, neither CO2-generating nor CO2-consuming buffers mitigated increases in coronary vein PCO2 or improved the outcome of cardiopulmonary resuscitation under these experimental conditions.

A comparison of venous blood gases during cardiac arrest

Previous reports have advocated the use of mixed venous blood gases to estimate arterial pH and as a reflection of tissue acid-based balance. However, true mixed venous samples are difficult to obtain during cardiac arrest as they require a pulmonary artery catheter. The purpose of this study was to determine whether central or femoral venous samples could be used in place of pulmonary artery samples. Blood gases from these sites were drawn at intervals during experimental cardiac arrest in dogs. The PO2, PCO2, and pH from the pulmonary artery samples were strongly correlated with those from the central venous (r = .93, .99, and .99, respectively) and from the femoral venous samples (r = .73, .93, and .97, respectively). There were no significant differences in the pulmonary artery, central, or femoral venous gases. This animal model suggests that femoral and central venous samples mirror true mixed venous blood gases from the pulmonary artery and could be used in their place.

The effect of epinephrine on hemodynamics, acid-base status and potassium during spontaneous circulation and cardiopulmonary resuscitation

The effect of a bolus dose of epinephrine on hemodynamics, acid-base status and potassium during spontaneous circulation and cardiopulmonary resuscitation (CPR) was investigated in 24 pigs weighing 20-25 kg over a period of 10 min. In a study of 12 pigs in a stable hemodynamic condition, at the 1- and 2-min point after injection of epinephrine or saline the mean serum potassium concentration was significantly higher in the six animals given epinephrine (6.9 +/- 0.7 and 5.4 +/- 0.6 mmol/l, respectively) than in the six control animals (3.8 +/- 0.6 and 3.9 +/- 0.4 mmol/l, respectively). At the later points of observation (3, 4, 5 and 10 min after injection of either epinephrine or saline) no significant difference was found between the groups. Following 1 min of ventricular fibrillation 12 pigs were resuscitated by closed-chest CPR. Six of these animals received 45 micrograms/kg epinephrine (epinephrine group), the other six animals were given physiological saline (control group). Mean aortic diastolic pressure during the relaxation phase was significantly higher in the epinephrine group than in the control group. There was no difference in cardiac index or acid-base status between the groups. In the epinephrine group mean arterial serum potassium concentrations reached a peak value of 6.7 +/- 1.1 mmol/l at 3 min after injection, when they were significantly (P less than 0.05) higher than in the control group (4.4 +/- 0.5 mmol/l). At 5 and 10 min, the potassium levels sank to 5.9 +/- 0.9 and 5.6 +/- 0.8 mmol/l, respectively, in the epinephrine group, and were no longer significantly different from the control group.

The effect of resuscitation technique and pre-arrest state of oxygenation on blood-gas values during cardiopulmonary resuscitation in dogs

Large mongrel dogs were anesthetized, instrumented, and subjected to electrically induced ventricular fibrillation after breathing either 100% oxygen (O2) or 10% O2 and 90% nitrogen for 10 minutes before arrest. Four minutes after arrest, open chest cardiopulmonary resuscitation (CPR) or intermittent abdominal compression closed chest CPR was initiated and continued for 20 minutes, at which time defibrillation was attempted. Central arterial and mixed venous blood samples were collected serially for the measurement of pH, carbon dioxide partial pressure (PCO2), and O2 partial pressure (PO2), and calculation of bicarbonate concentration and base excess. Mixed venous blood was collected serially for the measurement of lactate concentration. Hemodynamically variable resuscitation techniques and pre-arrest hypoxia or hyperoxia did not significantly influence blood-gas values during CPR. Mixed venous lactate concentrations after 20 minutes of CPR were significantly higher when hypoxia preceded the arrest and when intermittent abdominal compression closed chest CPR was used for resuscitation. Mixed venous PCO2 was significantly higher than arterial PCO2 in all dogs during CPR but was not significantly different before arrest.

Selective venous hypercarbia during human CPR: implications regarding blood flow

Thirty-five patients presenting to the emergency department in cardiopulmonary arrest had simultaneous measurement of central venous (cv) and arterial (a) blood gases during CPR with a pneumatic chest compressor and ventilator. The mean cv, arterial pH, and PCO2 values were markedly different (P less than .001). The mean pH gradient (pHa - pHcv) was .31 +/- .10 units and the mean PCO2 gradient (PcvCO2 - PaCO2) was 60.5 +/- 23.6 torr. This selective venous hypercarbia is probably due to a cardiac output that is inadequate to eliminate the CO2 produced from both residual aerobic metabolism and the buffering of anaerobically produced lactic acid. Central venous blood gases are probably a better reflection of actual tissue environment during prolonged cardiac arrest than are arterial blood gases.